PSI - Issue 68

Steffen Gerke et al. / Procedia Structural Integrity 68 (2025) 1294–1300 Gerke et al. / Structural Integrity Procedia 00 (2024) 000–000

1297

4

(a)

(d)

86

A

A

160

6

20

68

(c)

66

6

(b)

2

4

R3

R2

Fig. 1: (a) Specimen geometry; (b) cross section; (c) detail of notched part and (d) mesh of notched part, all measures in [mm].

case presented in (a) a maximum relative displacement of ∆ u max ref = 0

. 5 mm and a minimum relative displacement of

∆ u min ref = − 0 . 5 mm are reported resulting in failure of the specimen after 2 cycles where a cycle includes loading and reverse loading. For the second load case with constant maximum loads ∆ u max ref = 0 . 28mm and ∆ u min ref = − 0 . 25mm are measured which leads to failure after 6.5 cycles under compressive loading. A di ff erent approach was used for the experiments shown in Fig. 2(c, d): here the maximum and the minimum relative displacement have been increased in every cycle. For instance in (c) the displacement increase ∆ u inc ref = 0 . 15 mm has been applied leading to fracture after 3 cycles at ∆ u max ref = 0 . 58 mm. For the load case shown in (d) ∆ u inc ref = 0 . 05 mm has been used and ∆ u max ref = 0 . 48mm was reached whereby the test specimen failed under compressive loading. At this point, it is emphasized that the load cases shown in Fig. 2(a, c) failed under tension and that the load cases in Fig. 2(b, d) failed under compression. The corresponding numerically calculated load-displacement curves over all reflect nicely the experimentally obtained ones whereas under constant maximum loads the minimum force is slightly overestimated whereas for the load cases with load increase the maximum load is slightly underestimated. The strains on the specimen surfaces have been evaluated by digital image correlation (DIC) and by matching the relative displacements ∆ u ref of the numerical simulations the corresponding simulation results can be assigned. Fig. 3 displays the experimentally and numerically obtained first principal strain A 1 shortly before fracture. The results are displayed for the region marked by a red box in Fig. 1(a) and a suitable scale was selected for each load case. Overall, good agreement between experimental (EXP) and numerical (SIM) results can be recognized. The experiments that failed under tension loading (a,c) indicate a slightly to the left inclined shear band with maximum values of (a) 0.25 and (c) 0.35 which results from the more elevated relative displacement at the end of experiment (c), see Fig. 2(a, c). The experiment with 6.5 cycles and constant maximum loading (b) fractured under compressive loading and indicates a sharp band of elevated strains with values up to 0.13 and is inclined to the right. The experimental data (b, left) is more noisy and consequently the sharp characteristic is less visible. The experiment with 8.5 cycles and load increase indicates a to the left inclined band with elevated strains. Unlike in (b), the maximum values (0.17) are not continuous across the height but are ’flame-shaped’ towards the center. The experimental data is not as well resolved here either. Thus, the influences of the load history is clearly visible, which is additionally emphasized by slightly increased strains outside the band. Consequently, the loading direction as well as the loading history influences the inclination and occurrence of the shear band. The accumulated principal damage strains A da ( i ) based on Eq. (4) are accessible through the numerical simulations. Fig. 4 shows the calculated first principal damage strain A da 1 for all four load cases shortly before fracture on the left